A Novel Approach to the Part Orientation Problem for Robotic Asse

Marquette University e-Publications@Marquette Master's Theses (2009 -) Dissertations, Theses, and Professional Projects A Novel Approach to the Part Orientation Problem for Robotic Assembly Applications Brian James Slaboch Marquette University Recommended Citation Slaboch, Brian James, "A Novel Approach to the Part Orientation Problem for Robotic Assembly Applications" (2011) Master's Theses (2009 -) Paper 73 http://epublications.marquette.edu/theses_open/73 A NOVEL APPROACH TO THE PART ORIENTATION PROBLEM FOR ROBOTIC ASSEMBLY APPLICATIONS by Brian J Slaboch, B.S A Thesis Submitted to the Faculty of the Graduate School, Marquette University, in Partial Fulfillment of the Requirements for the Degree of Master of Science Milwaukee, Wisconsin May 2011 ABSTRACT A NOVEL APPROACH TO THE PART ORIENTATION PROBLEM FOR ROBOTIC ASSEMBLY APPLICATIONS Brian J Slaboch, B.S Marquette University, 2011 SCARA (Selective Compliant Assembly Robot Arm) type robots are the most common type of assembly robots These robots have four degrees of freedom (three rotational and one translational) Typically these robots are used for assembly tasks that take place along a vertical axis Many times, however, assembly tasks take place along a non-vertical axis To account for non-vertical axis assembly, parts must be fed in a proper orientation to allow for correct assembly Parts feeders and specialized end-effectors are typically used to feed parts in their proper orientation This thesis investigates a novel end-effector that can be used to feed parts for industrial assembly applications Specifically, the purpose of the novel end-effector is to provide a SCARA robot with an added selectable degree of freedom This end-effector aims to bridge the gap between complex anthropomorphic grippers and simple binary grippers The approach is novel in that the end-effector interacts with the environment to produce the added degree of freedom New path planning algorithms were developed to work in conjunction with the novel end-effector A prototype end-effector was designed, built, and tested to prove the validity of this new approach i ACKNOWLEDGEMENTS Brian J Slaboch, B.S I would first and foremost like to thank Dr Philip A Voglewede for providing me with this great opportunity This work could not have been completed without his support and guidance In addition, I would like to thank Dr Mark Nagurka and Dr Joseph Schimmels for their insightful comments and suggestions A special thanks goes to Jinming Sun and Bryan Bergelin for their continued support and helpful advice I would also like to thank Tom Silman and Ray Hamilton for their work in the machine shop Finally, the I would like to thank my family and friends for their encouragement and guidance ii DEDICATION Brian J Slaboch, B.S I would like to dedicate this work to my family and friends Without their love and support none of this would have been possible Thank You iii TABLE OF CONTENTS ACKNOWLEDGEMENTS i DEDICATION ii TABLE OF CONTENTS iii LIST OF TABLES v LIST OF FIGURES vi CHAPTER Introduction 1.1 Parts Feeding Systems 1.2 End-Effector Design 1.3 Summary CHAPTER Mechanical Design 10 2.1 Design Requirements 10 2.2 Engineering Requirements 11 2.3 Quality Function Deployment 13 2.4 Conceptual Design 13 2.4.1 Magnet Device 13 2.4.2 Ratchet Device 14 2.4.3 Mechanical Brake 14 2.4.4 Friction Device 14 2.5 Concept Selection 15 2.6 Configuration Design 17 2.7 Parametric Design 19 CHAPTER Path Planning 22 3.1 Path Planning 22 3.1.1 Horizontal Line Path 23 3.1.2 45◦ Angle Path 24 3.1.3 Shortest Distance Path 25 3.2 Path Planning Algorithms 26 3.2.1 Horizontal Line Path Algorithm 27 3.2.2 45◦ Path Algorithm 29 3.2.3 Shortest Distance Path Algorithm 32 3.2.4 Return Path 34 CHAPTER Dynamic Analysis 37 4.1 Lagrange’s Equations of Motion 37 4.2 Newton-Euler Equations of Motion 42 4.2.1 Vertical Motion 42 4.2.2 Rotation about the θ3 Axis 43 4.2.3 Rotation about the θ1 Axis 45 iv TABLE OF CONTENTS — Continued CHAPTER End-Effector Manufacturing and Testing 5.1 Material Selection 5.2 Positioning Hinge Selection 5.3 Detail Design 5.4 Prototype Testing 5.4.1 Rapid Prototype Test 5.4.2 Rapid Prototype Test 5.4.3 Design Modifications 5.5 Final Design 5.5.1 Final Design Test 5.5.2 Final Design Test 5.5.3 Discussion CHAPTER Contribution and Future Work 6.1 Contributions of this Research 6.2 Future Work 6.2.1 Dynamic Analysis 6.2.2 Kinematic Analysis 6.2.3 Stress Analysis 6.2.4 Robustness REFERENCES APPENDIX A 50 50 51 53 54 56 57 58 58 59 59 60 62 62 63 63 64 65 65 66 68 v LIST OF TABLES 2.1 2.2 2.3 2.4 4.1 5.1 Pairwise Comparison Engineering Requirements Weighted Rating Method Scale Maximum Accelerations and Costs Velocities 11 12 16 16 48 53 vi LIST OF FIGURES 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.1 2.2 2.3 2.4 2.5 2.6 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11 3.12 3.13 3.14 3.15 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 5.1 Manipulation Flow Chart SCARA ROBOT Lever Assembly onto Vertical Post [1] Angled Peg Assembly using Adept Viper Robot Utah/Mit Hand [2] Pivot Grasp [3] Schunk SKE Pneumatic Swivel Head Metamorphic Gripper Underactuated Robot Magnet Concept Concept Sketch: Friction Device Configuration Configuration Configuration Critical Dimension Pivoting Gripper Device Schematic Horizontal Line Path Angle Path 90◦ Pick and Place [4] Shortest Distance Path Simplified Geometry Horizontal Line Path Schematic θf vs h (L = 8.9 cm) θf vs h 45◦ Angle Schematic d vs θf αmin Shortest Distance Path Schematic Return Path Schematic Return Path SCARA Robot with End-Effector Reference Configuration Free Body Diagram, Vertical Motion Case Free Body Diagram, Rotation about the θ3 Axis τ vs θ4 for θ3 rotation Free Body Diagram, Rotation about the θ1 Axis τ vs θ4 for θ1 rotation (θ3 = 90◦ ) τ vs θ4 for θ1 rotation (θ3 = 0◦ ) τ vs θ4 for Horizontal Motion Reell PHK Positioning Hinge 3 8 14 17 18 18 19 20 23 24 25 25 26 27 28 29 30 30 31 33 33 34 35 38 40 41 44 44 45 47 47 48 52 vii 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9 5.10 5.11 A.1 A.2 LIST OF FIGURES Detailed Drawing of the Pivot Arm Detailed Drawing of the Base CAD Model Rapid Prototype Test with Rapid Prototype Test with Rapid Prototype Final Design Final Design with Gripper Final Design Test Final Design Test 45◦ Angle Schematic Relation Between ǫ and l — Continued 54 55 55 56 57 58 59 60 60 61 68 69 55 166± 010 (#19 drill for M5x0.8 tap) 1.299±.010 1.25 95 R.25 1.18 2.36±.01 1.882±.010 24±.01 1.05±.01 2.65±.01 375±.010 190 R1.00 R.25 4.38 38±.01 R.125 25 159± 010 75±.01 (#21 drill for 10-32 tap) 1.40 375 25±.01 R.25 50±.01 1.00±.01 MARQUETTE UNIVERSITY SCALE: 0.5 6-10-2010 SIZE: A WEDE LAB AL 7075-T6 DRAWN BY: BASE BRIAN SLABOCH Figure 5.3: Detailed Drawing of the Base Figure 5.4: CAD Model Having a physical prototype ensured that the parts could be assembled properly In addition, it provided a reality check on the size of the components After examining the prototype it was determined that the fillets could have a larger radius to reduce stress Ideally, this device would be tested on a SCARA type robot However, the only available robot for testing was a Mistsubishi Melfa RV-3S six-axis robot To 56 mimic the motion of a SCARA type robot the six DOF Mitsubishi robot was constrained to move with only four DOF Base Positioning Hinge Rapid Prototype Gripper Pivot Arm Fixed Post Figure 5.5: Rapid Prototype 5.4.1 Rapid Prototype Test The first test was a ninety degree turn test shown in Fig 5.6 This test is meant to mimic common pick and place assembly operations For this test the pivot arm was placed next to the fixed post as shown in Fig 5.6a The 45◦ path algorithm was used to pivot the gripper 90◦ as shown in Fig 5.6b The speed on the robot was set to 30% of its maximum value The results of this test were encouraging The gripper was successfully cycled through this process fifty times Visual results showed that the gripper was routinely pivoted to approximately the same place after each cycle was completed 57 (a) (b) Fixed Post Figure 5.6: Test with Rapid Prototype 5.4.2 Rapid Prototype Test The second test performed was completed to determine the relative accuracy of the device The end-effector was used to place a knife in an out of an angled knife holder This is shown in Fig 5.7 The robot was set at 10% of it maximum speed The horizontal path algorithm was used to rotate the knife to the same angle as the slot in the knife holder The knife was then placed into the knife holder The knife was then pulled out of the knife holder, and the gripper was returned to its initial position This cycle was repeated 25 times This shows that the device is repeatable The knife has a thickness of 1.2 mm and the slot it was placed into has a thickness of 4.42 mm Thus, the prototype had a positioning error of ±1.61 mm for this particular test After the program was run at 10% of the robot maximum speed, the speed was increased to 20% Increasing the speed of the robot to 20% did not effect the results The knife was still accurately placed in and out of the knife holder However, this was not the case when the speed of the robot was further increased to 30% At this higher speed the impact forces caused the pivot arm to rotate slightly more than desired This means that the knife was not able to be correctly placed into the knife holder As the knife approached the slot the tip of the knife blade caught onto the solid part of the knife holder 58 (a) (b) Figure 5.7: Test with Rapid Prototype 5.4.3 Design Modifications From these tests multiple lessons were learned The first lesson learned was that the fixed post was not sturdy enough In these tests a hollow metal tube was clamped to a wood board When the pivoting part was pressed through the fixed post, the post had a small vibration A studier fixed post was used for the final design In Test an important lesson was that the impact forces are high enough that at different robot speeds the gripper will be rotated a different amount This relates back to the path planning algorithms developed in Chapter As stated previously, these algorithms are developed based solely on the kinematics and not on the dynamics of the motion In practice, the dynamics cannot be ignored At very low speeds the impact forces can essentially be ignored Therefore, it is recommended that the path planning algorithms be used at lower speeds However, as will be seen in future sections, it is still possible to effectively use this device at higher speeds 5.5 Final Design The final product is shown in Fig 5.8 Multiple tests were performed to determine the effectiveness of the design This device was designed to be attached 59 to a Schunk PWG-60s angular two finger binary gripper Figure 5.9 shows the gripper attached to the end-effector (a) (b) (c) Figure 5.8: Final Design 5.5.1 Final Design Test Figure 5.10 shows the setup for Test 12 In this demonstration a block is picked up from the base plate of the robot, rotated to coincide with the angle of the ramp, and finally the block is placed on the ramp and it slides back down to the base plate of the robot This cycle is then repeated multiple times The robot speed was set to the 40% of the maximum value During this test it was seen that the block was continuously brought to the correct angle, and there was almost no discrepancy between cycles 5.5.2 Final Design Test Figure 5.11 shows a 90◦ turn test that was performed This is similar to the test that was performed with the prototype gripper The difference was that the speed was set to the maximum allowable speed of the robot to prove that this device would work at high speeds The device was put through approximately 200 cycles The average cycle time was 1.8 s This fell within the specified engineering requirements This demonstration was completed with help from Kevin Flynn 60 (a) (b) Figure 5.9: Final Design with Gripper Figure 5.10: Final Design Test 5.5.3 Discussion There were multiple lessons learned during the testing of the end-effector As stated previously, a sturdier fixed post was needed Fig 5.10 shows the modified version of the fixed post The new version includes a solid rod that is bolted to a stand that is clamped to the table of the robot It was found that with this newly designed fixed post, there was less vibration upon impact For laboratory testing the fixed post is simply clamped onto the stationary base plate of the robot One reason for this is that it allows for the capability to move the fixed post to any location along the base plate In practice, however, it is recommended that the fixed post be mounted directly to the base plate of the robot without using any clamps 61 (b) (a) Figure 5.11: Final Design Test This will allow for a sturdier and more accurate system All of the tests were run using the teach-repeat option for the robot One drawback that occurred in the 90◦ turn test (max speed case) was that the robot had to be programmed at its max speed The program could not be created at a lower speed and then be modified to run at the max speed The reason for this, as mentioned previously, was that due to the high impact forces the pivot arm would overshoot it desired angle This meant that the programmer had to correct for this offset The offset was significant enough that simply using the kinematic path planning from Chapter would not produce the desired result The positive result from this was that it was relatively easy to program the robot to move the gripper to the desired angle The kinematic path planning algorithms allowed one to determine approximately where the end-effector should be moved to From there it was simple to correct for any additional springback from the high impact forces Ideally, the motion planning algorithm will incorporate the dynamics This will be an area of future work In this chapter material selection, positioning hinge selection, detail design, and experimental testing was discussed The results from the experimental tests show that this gripper is industrially feasible, and it did meet the design requirements The final product is a low weight gripper that provides a high torque to weight ratio 62 CHAPTER Contribution and Future Work Orienting parts properly for robotic assembly applications is a difficult problem Researchers have developed ways to orient parts ranging from parts feeders to complex anthropomorphic hands Each of these solutions works well for specific types of robotic assembly applications This thesis focused on a particular subset of robotic assembly tasks The goal of this research was to develop an end-effector that could be used in an industrial setting to provide a selectable DOF to a SCARA type robot The next section outlines the specific contributions of this research 6.1 Contributions of this Research The contributions of this research are as follows: • A novel end-effector was designed, built, and tested • The flexibility in this design is a key contribution Many grippers that are aimed at providing an added DOF only work for polyhedral parts This design is built to work with any custom gripper that can pick up a wider range of parts • New path planning techniques were developed to be used with the end-effector Most path planning techniques are aimed at collision avoidance This is an application in which obstacles can be used as an advantage • This end-effector shows how the built in controls of a robot can be used to aid in part manipulation The end-effector can be designed as part of the overall system as opposed to a separate entity • A dynamic analysis was completed to determine the required resistance torque for the positioning hinge The closed form solution allows for designers to 63 easily specify an appropriate positioning hinge for a particular application It is to be emphasized that part of the beauty in this design is in its simplicity A simple mechanical device can be used to replace a complex electrical or pneumatic rotary actuator This end-effector is not a panacea for all robotic assembly applications However, this is a step in the right direction The next section will explore future research opportunities based on this work 6.2 Future Work There is a considerable amount of future work that can be done based on this research There are many questions that need to be answered before this work can be implemented in an industrial setting 6.2.1 Dynamic Analysis In this thesis a dynamic analysis was completed that enables a designer to determine the necessary torque the positioning hinge must provide to ensure the positioning hinge does not move due to either inertial forces or gravity However, there is more dynamic analysis that needs to be completed Most importantly, there are significant impact forces that occur when the pivot arm makes contact with the fixed post at high speeds If the impact forces are too high the positioning hinge will rotate more than the desired angle of rotation determined from the kinematic modeling Therefore, impact forces should be reduced as much as possible One way to reduce impact forces would be to coat the fixed post with rubber to help absorb impact forces Also, the robot’s end-effector may be moved at a slow enough speed that impact forces can be ignored The impact forces could be used to determine the desired angle of rotation This could be determined experimentally For horizontal path motion the angle of rotation, θf , could be a function of multiple variables as shown in Eq 6.1 θf = f (mp , mg , vh , h, τ ), (6.1) 64 where mp and mg are the mass of the part and the gripper, vh is the horizontal velocity of the end-effector, h is the distance from the centerline of the positioning hinge to the center of the fixed post, and τ is the amount of torque provided by the positioning hinge This model could be used in conjunction with the kinematic analysis in this thesis to create a more accurate system Another dynamic problem of interest is the amount of overshoot that may occur as the gripper is being rotated The momentum from the gripper may cause overshoot This can be determined by using a standard dynamic analysis 6.2.2 Kinematic Analysis This thesis outlined three related path planning algorithms that may be used to obtain the desired angle of rotation In each of these path planning algorithms the following assumptions were made: Only one fixed post was used The fixed post was cylindrical The pivot arm was a flat surface With these assumptions in place three path planning algorithms were developed Each of the path planning algorithms involved moving the end-effector in a straight line, and the algorithms are all related in a general way Only straight line paths were considered because moving the end-effector in a straight line reduced the cycle time However, there are still kinematic questions that need to be answered A more comprehensive way to approach the kinematic problem would be to approach the problem from a complete system perspective The goal of the kinematic problem is the complete the following steps in the most efficient and accurate way possible: Grasp the part Rotate the gripper from its initial angle to a desired angle Assemble the part 65 Rotate the gripper back to its initial angle Repeat this process When viewing this problem from a system perspective the problem becomes more complex There are new questions that would need to be answered For instance, where should the fixed post be placed in the robot’s workspace to provide optimal performance? Additionally, there may be advantages to using more than one fixed post By strategically placing fixed post’s in the robot’s workspace an optimal solution could be obtained Another consideration that would need to be taken into account is the design of both the fixed post and the pivot arm The kinematic equations developed in this thesis were dependent on the radius, r, of the fixed post Thus, the design of both the fixed post and the pivot arm critical in the path planning 6.2.3 Stress Analysis A standard stress analysis could be completed on the fixed post to ensure that it can withstand the continuous impact forces that occur Additionally, a standard stress analysis could be completed on the pivot arm to ensure that it does not yield to the continuous impact forces 6.2.4 Robustness The most important part of a robotic system is robustness In this design a positioning hinge was used that will work for 20,000 cycles More research needs to be completed to determine a way to create more robust positioning hinges that will last for longer periods of time Additionally, over time there may be wear on both the pivot arm and the fixed post Research needs to be done to determine how much of an impact wear will have on the overall system 66 REFERENCES [1] M.T Zhang and K Goldberg Gripper point contacts for part alignment IEEE Transactions on Robotics and Automation, 18(6):902–910, December 2002 [2] S Jacobsen, J Wood, K Bigger, and E Iverson The Utah/MIT hand: Work in progress International Journal of Robotics Research, 4(3):21–50, 1984 [3] B Carlisle, K Goldberg, A Rao, and J Wiegley A pivoting gripper for feeding industrial parts In Proceedings of the 1994 IEEE International Conference on Robotics and Automation, volume 2, pages 1650–1655, San Diego, CA, May 1994 [4] A Rao, D.J Kriegman, and K Goldberg Complete algorithms for feeding polyhedral parts using pivot grasps In IEEE Transactions on Robotics and Automation, volume 12, pages 331–342, April 1996 [5] G Boothroyd, C Poli, and L.E Murch Automatic Assembly Marcel Dekker, Inc., New York, 1982 [6] G Boothroyd, P Dewhurst, and W Knight Product Design for Manufacture and Assembly Marcel Dekker, Inc., New York, 1994 [7] M A Peshkin and A.C Sanderson Planning robotic manipulation strategies for workpieces that slide IEEE Transactions on Robotics and Automation, 4(5):696–700, October 1988 [8] M.T Zhang, K Goldberg, G Smith, R.-P Berretty, and M Overmars Pin design for part feeding Robotica, 19(6):695–702, November/December 2001 [9] G Causey and R Quinn Design of a flexible part feeding system In Proceedings of IEEE International Conference on Robotics and Automation, Albuquerque, NM, 1997 [10] M.T Mason and J.K Salisbury Robot Hands and the Mechanics of Manipulation MIT Press, Cambridge, MA, 1985 67 [11] William T Townsend The BarrettHand grasper programmably flexible part handling and assembly Industrial Robot: An International Journal, 27(3):181–188, 2000 [12] Jiansheng Dai and Zhang Qixian Metamorphic mechanisms and their configuration models Chinese Journal of Mechanical Engineering, 13(3):212–218, 2000 [13] Jacob A Ziesmer and Philip A Voglewede Design, analysis and testing of a metamorphic gripper In Proceedings of the ASME 2009 International DETC, 2009 DETC2009-87512 [14] K.M Lynch and M.T Mason Dynamic manipulation with a one joint robot In Proceedings of the 1997 IEEE International Conference on Robotics and Automation, pages 359–366, Albuquerque, NM, April 1997 [15] K.M Lynch, N Shiroma, H Arai, and K Tanie Collision-free trajectory planning for a 3-dof robot with a passive joint The International Journal of Robotics Research, 19(12):1171–1184, December 2000 [16] R Eggert, editor Engineering Design Pearson Prentice Hall, Upper Saddle River, NJ, 1999 [17] J Latombe Robot Motion Planning Kluwer Academic Publishers, sixth edition, 2001 [18] J Ginsberg Mechanical and Structural Vibrations: Theory and Applications John Wiley and Sons, 2001 68 APPENDIX A The derivation below is for the distance, d, from (xi , yi ) in terms of h, R, and θf as shown in Fig A.1 R ε L θf h B β p xf , yf d α A xi , yi Figure A.1: 45◦ Angle Schematic From Fig A.1 ǫ = π − θf (A.1) β = π − α − θf (A.2) From Fig A.2 ǫ R = l sin (A.3) Substituting Eqn A.1 into Eqn A.3 gives R = l cos θf (A.4) π−ǫ , (A.5) The distance p can be written as p = h − l sin 69 Figure A.2: Relation Between ǫ and l which can be simplified to p = h − l sin θf (A.6) Substituting the relation between R and l from Eqn A.4 gives p=h− R sin cos θf θf (A.7) The law of sines states that sin θf sin β = d p (A.8) Substituting β and p into Eqn A.8 and solving for d gives d= sin θf sin (π − α − θf ) h − Rtan θf (A.9) .. .A NOVEL APPROACH TO THE PART ORIENTATION PROBLEM FOR ROBOTIC ASSEMBLY APPLICATIONS by Brian J Slaboch, B.S A Thesis Submitted to the Faculty of the Graduate School, Marquette University, in Partial... rotational and one translational) Typically these robots are used for assembly tasks that take place along a vertical axis Many times, however, assembly tasks take place along a non-vertical axis... that of the robot Motors may lead to significant downtime and added cost Another option is to attach a pneumatic rotary actuator to the end of a SCARA type robot to provide an added DOF An example